Saturday, 2 December 2023

A Guide to Python NLP Topic Extraction

16 Feb 2023

Natural Language Processing (NLP) is a field of computer science, artificial intelligence, and computational linguistics concerned with the interaction between computers and human language. With the ever-growing amount of text data generated every day, extracting meaningful insights from such data has become a critical task for businesses, researchers, and individuals. One of the most common tasks in NLP is topic extraction, which involves identifying the main themes or topics present in a given text.

Python is one of the most popular programming languages for NLP, thanks to its rich ecosystem of libraries and tools for working with text data. In this guide, we will explore some of the best Python libraries for topic extraction and provide a step-by-step tutorial on how to use them to extract topics from text.

Introduction to Topic Extraction

Topic extraction is the process of automatically identifying the main themes or topics present in a given text. This task is often used to summarize large amounts of text data or to understand the main ideas present in a document or corpus.

There are many different approaches to topic extraction, including unsupervised methods like Latent Dirichlet Allocation (LDA) and Non-negative Matrix Factorization (NMF), and supervised methods like Support Vector Machines (SVMs) and Neural Networks. In this guide, we will focus on unsupervised methods, which are widely used for topic extraction and do not require labeled data.

Python Libraries for Topic Extraction

There are several Python libraries available for topic extraction, each with its strengths and weaknesses. Here are some of the most popular libraries:


Gensim is a Python library for topic modeling and document similarity. It provides a simple interface for training topic models using popular algorithms like LDA and LSI (Latent Semantic Indexing) and for extracting topics from new documents.


spaCy is a Python library for advanced NLP tasks like named entity recognition, dependency parsing, and part-of-speech tagging. It also includes a built-in functionality for topic modeling, which can be used to extract topics from a corpus of documents.Scikit-learn

Scikit-learn is a popular machine learning library in Python. It includes a wide range of algorithms for classification, regression, clustering, and other tasks. It also includes a module for topic modeling, which provides implementations of algorithms like LDA and NMF.

Topic Extraction with Gensim

In this section, we will provide a step-by-step tutorial on how to extract topics from a corpus of documents using Gensim.

Step 1: Load Data

The first step is to load the data into Python. For this example, we will use a dataset of news articles from the Reuters corpus, which is included in the NLTK library.

import nltk

from nltk.corpus import reuters

# Load the data

documents = reuters.fileids()

Step 2: Preprocess Data

Next, we need to preprocess the data by tokenizing the documents, removing stop words, and stemming the remaining words.

from nltk.tokenize import word_tokenize

from nltk.corpus import stopwords

from nltk.stem import PorterStemmer

# Tokenize the documents

tokenized_docs = [word_tokenize(reuters.raw(doc_id)) for doc_id in documents]

# Remove stop words

stop_words = set(stopwords.words(‘english’))

filtered_docs = [[word for word in doc if word.lower() not in stop_words] for doc in tokenized_docs]

# Stem the words

stemmer = PorterStemmer()

stemmed_docs = [[stemmer.stem(word) for word in doc] for doc in filtered_docs]

Step 3: Create Dictionary and Corpus

Next, we need to create a dictionary and corpus from the preprocessed documents. The dictionary maps each word in the corpus to a unique integer ID, and the corpus is a list of bag-of-words representations of each document.

from gensim.corpora import Dictionary

from gensim.models import TfidfModel

# Create dictionary

dictionary = Dictionary(stemmed_docs)

# Create corpus

corpus = [dictionary.doc2bow(doc) for doc in stemmed_docs]

# Create TF-IDF model

tfidf = TfidfModel(corpus)

corpus_tfidf = tfidf[corpus]

Step 4: Train LDA Model

We are now ready to train our LDA model. We will use the Gensim implementation of LDA, which requires us to specify the number of topics we want to extract.

from gensim.models import LdaModel

# Train LDA model

num_topics = 10

lda_model = LdaModel(corpus_tfidf, num_topics=num_topics, id2word=dictionary, passes=10)

Step 5: Extract Topics

Finally, we can extract topics from our corpus using the trained LDA model. The print_topics() method of the model prints the most probable words for each topic.

# Extract topics

topics = lda_model.print_topics()

# Print topics

for topic in topics:



In this guide, we have explored the topic of topic extraction in NLP and provided a step-by-step tutorial on how to extract topics from text using the Gensim library in Python. We have also discussed other Python libraries for topic extraction, including spaCy and Scikit-learn, and their strengths and weaknesses.

Topic extraction is a crucial task in NLP, and Python provides a rich ecosystem of libraries and tools for working with text data. With the help of these libraries, businesses, researchers, and individuals can extract valuable insights from large amounts of text data and gain a deeper understanding of the main themes and ideas present in a document or corpus.